Figure 8.

The infection scale (top row) and cumulative infection scale (bottom row) on four networks with different fractions of nodes removed under various strategies, where $\beta =1.5$, $\gamma =1$ and $p=0.01$. In the top row, JN performs better than FN in all the fractions and all the datasets, since JN generates lower and postponed infection scale peaks compared to FN. With the increasing of the nodes removal fraction, the advantage of JN grows. In the bottom row, immunizing the same number of nodes identified by JN can reduce the network’s connectivity more than FN, resulting in a smaller final infection scale. For example, the final infection scale decreases from 40 to 32% when immunizing 30% nodes identified by JN compared with FN in SYN-BAc network. The results are averaged over 20 independent runs.